3. Deduct product classes

Whereas the construction of the dendrogram is based on increasing dissimilarity or aggregation indices, the use of the dendrogram proposed here consists in going down the dendrogram branch by branch, class by class, to deduce "product classes" which serve as product models for the comparison of replaced and innovative modules. The control criterion is the relative error committed by the class on each product of the same class; the maximum relative error can initially be set at 10%. As soon as the relative error committed by the class falls below the 10% threshold, the branch is stopped. Otherwise, the class is divided into two sub-classes (depending on the branching of the dendrogram) and the process continues. Taking the previous example, we can see that four product classes have been created (see figure Example of product class deduction in a dendrogram).

Once...